Preparation of γ-Al2O3/α-Al2O3 ceramic foams as catalyst carriers via the replica technique

Article abstract of DOI:10.1016/j.cattod.2020.09.019

This work describes an effective method for the preparation of open-cell ceramic foams for their further use as catalyst supports. The polyurethane sponge replica technique was applied using a ceramic suspension based on a mixture of α-alumina, magnesia a

Full text of DOI:10.1016/j.cattod.2020.09.019

Catalysis Today xxx (xxxx) xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Preparation of γ-Al2O3/
α
-Al2O3 ceramic foams as catalyst carriers via the
replica technique
Vladimir Shumilova, Alexey Kirilina, Anton Tokareva, Stephan Bodenb, Markus Schubertb,
Uwe Hampelb, Leena Hupaa, Tapio Salmia, Dmitry Yu. Murzina *
,
a Johan Gadolin Process Chemistry Centre, Abo Akademi University, FI-20500 Turku/Abo, Finland
b Institute of Fluid Dynamics, Helmholtz-Zentrum Dresden-Rossendorf, Bautzner Landstr. 400, 01328 Dresden, Germany
A R T I C L E I N F O
A B S T R A C T
Keywords:
This work describes an effective method for the preparation of open-cell ceramic foams for their further use as
catalyst supports. The polyurethane sponge replica technique was applied using a ceramic suspension based on a
γ-Al2O3/α- Al2O3
Catalytic foams
mixture of α-alumina, magnesia and titania and polyvinyl alcohol solution as a liquid component. The poly-
Macroporous ceramics
urethane sponge was etched with NaOH and covered with colloidal silica to obtain better adhesion of the slurry
to the walls of the polymeric material onto it. The surface area of the ceramic carrier was increased by adding a
layer of γ-alumina. Deposition of an active catalytic phase (Pt) was done by impregnation. Properties of the
carriers and the final catalyst were investigated by a number of physico-chemical methods such as TEM, SEM,
XRD and computer tomography. Hydrogenation of ethyl benzoylformate was performed to elucidate the catalytic
properties of foam catalysts illustrating their applicability.
Hydrogenation of ethylbenzoylformate
1. Introduction
shock resistance. Therefore, such foams are commonly applied for
thermal insulation and fire protection materials, while the open-cell
One of the current important topics in catalytic reaction engineering
is the optimization and intensification of catalytic reactors with the aid
of advanced structured materials. For three phase processes structural
packing are used to enhance the gas-liquid-solid contact area and tur-
bulence within the fluid phase, thus increasing the interfacial mass
transfer. It has been demonstrated for two- phase and three-phase re-
actions that structured packings, e.g. open cell foams, have better hy-
drodynamic performance than conventional packings such as for
example Raschig rings [1]. Cellular structures perform remarkably well
in continuous liquid- phase catalytic processes giving a low pressure
drop and enhanced mass transfer [2].
ceramic foams are particularly used for molten metal filtration, diesel
engine exhaust filters and hot gas filtration [6,8]. Free space between
the ceramic particles allows a large pore volume and a porosity ranging
from nanometers to microns or even larger. Subsequently, ceramic
foams can be used as membranes, absorbents, in insulation and
biomedical devices, as well as catalytic converters [9–11].
Ceramic foam can be produced using various methods, including the
replica method [12], starch consolidation [13], the foaming method
[14] and gel-casting [11].
The replica method allows an open pore volume of up to 95 % giving
an open network of interconnected, but hollow struts. The structure of
the resulting foam can be either random or regular depending on the
template. Obviously, a large pore volume is accompanied with a low
pressure drop and density.
Macroporous ceramic foams are porous, lightweight, solid materials
with high density, large specific surface area, high thermal stability and
resistance to chemical attacks [3,4]. Ceramic foams consist of cellular
structures composed of three-dimensional networks of struts [5].
The common types of ceramic foams are made of silicon carbide,
alumina, zirconia, titania, and silica [6,7]. Depending on the cellular
structure, the foams are categorized as either open or closed cell foams.
In the closed pores there is no or limited connection between the single
pores, which might be beneficial for applications requiring high thermal
In this work, the replica method was used, which involves coating of
the polymer sponge with the ceramic slurry followed by removal of the
excess slurry by squeezing, blowing with air to ensure permeability of
pores, drying and sintering to remove the polymer components [9]. This
procedure results in a ceramic replica of the original polymeric foam. A
common challenge in producing porous ceramics with the replica
* Corresponding author.
E-mail address: dmurzin@abo.fi (D.Yu. Murzin).
https://doi.org/10.1016/j.cattod.2020.09.019
Received 12 March 2020; Received in revised form 2 June 2020; Accepted 21 September 2020
Available online 30 September 2020
0920-5861/© 2020 Elsevier B.V. All rights reserved.
Please cite this article as: Vladimir Shumilov, Catalysis Today, https://doi.org/10.1016/j.cattod.2020.09.019

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Fig. 1. Heating schedules for calcination of α-alumina support a) per se and b) covered with γ-alumina layer.
method is the choice of a suitable pore size template while retaining an
adequate mechanical strength of the final foam. Optimization of
different parameters in preparation of the ceramic slurry, subsequent
coating of the sponge with the slurry, and sintering is needed to achieve
final foams with the desired properties. The sponge used as a replica
must be flexible and able to regain its original shape after immersion in
the slurry and squeezing. The choice of the polyurethane (PU) sponge
pore size and shape is defined by the requirements of the final foam
application.
which it was dried at 70 ◦C for 1 h. The samples were washed with water
and ethanol and then air-blown followed by two immersions in 30 wt.%
silica sol solution (30 wt.%) (Aldrich, 30 wt.%, CAS number
[7631ꢀ 86-9]). After each immersion, the samples were squeezed and
blown by air, and dried at 70 ◦C for two hours. Two immersions were
required to adhere a ceramic layer thick enough on the PU surface thus
aiding in subsequent coating with α-alumina. As a liquid component of
the slurry 5 wt.% PVA (Acros Organics, 5 wt.%, CAS number
[9002ꢀ 89-5]) solution was used.
As a carrier material
α-alumina was used in the current work,
The slurries were prepared using the following procedure: 9.8 g of
providing required thermal stability. The replica technique was utilized
to manufacture a macroporous support by coating the polymer sponge
α-alumina (Treibcher Schleifmittel/Alodur 220), 0.05 g of titania
(Aldrich, CAS number [1317ꢀ 70-0]) and 0.05 g of magnesia (Fluka,
CAS number [1309ꢀ 48-4]) were mixed with 4.5, 5 or 5.5. mL of PVA
solution. The resulting slurries were ground in the ball milling machine
(Philips MiniMill) for 45 min.
with a ceramic slurry containing α-alumina and a range of additives
acting as binders, rheological and anti-foaming agents [15].
The surface area of α-alumina is typically not sufficient to deposit a
catalytically active phase, therefore washcoating with γ-alumina was
required. Such procedure allows to substantially enhance the surface
area from few m2/g to ca. 250ꢀ 300 m2/g [16].
The coating process included immersion of the sponge in the slurry,
air-blowing to remove the excess of slurry and then drying at 70 ◦C.
Immersion was done two to three times. The PU sample might be
compressed before the first immersion, therefore it was allowed to
expand while being immersed in the slurry. In a typical experimental
series, the amount of the liquid phase was increased with every im-
mersion of the PU sample into the slurry in order to keep pore
permeability.
While there are few publications in the literature on catalytic foams
development discussing mechanical strength and porosity [17–19] there
is a lack of studies addressing the dependence of these parameters on the
preparation procedure. The intent of this article is to explore the impact
of the preparation procedure on the properties of catalytic foams. The
scope of the work is in particular on the investigation of the influence of
the number of ceramic layers on strength, porosity and pore inter-
connectivity of the foams.
Calcination was done in an electrically heated furnace. The coated
samples were calcined at 1500 ◦C for 3 h according to the heating
schedule shown in Fig. 1a. Thermal degradation polyurethane foams has
been investigated previously in the literature using thermogravimetry
[28]. From differential scanning calorimetry the maximum temperature
were ca. 315 and 530 ◦C, thus longer dwelling at two temperatures
corresponding to the main degradation stages was applied. The tem-
perature was slowly increased from room temperature to 1500 ◦C first
by heating at a rate 50 ◦C/h to 300 ◦C and then to 600 ◦C with a ramp
5 ◦C/min. Isothermal conditions were maintained for 1 h at 300 ◦C and
600 ◦C in order to burn out the PU matrix. Heating from 600 ◦C to
1500 ◦C was done with a ramp with a ramp 5 ◦C/min. Cooling down
from the top firing temperature was slow (5 ◦C/min) in order to prevent
formation of thermal stresses, which could induce cracks in the foam.
Application of ceramic alumina based foams in catalysis has been
discussed in the literature [20–22]. In particular, such foams were
extensively used in various hydrogenation reactions [23–25].
Hydrogenation of ethyl benzoylformate (EBF) [26] in a continuous
up-flow fixed bed reactor was selected as a test reaction in this work. If
this reaction is performed with structural catalysts, a catalytically active
phase (e.g. Pt) should be introduced on the ceramic foam. To this end the
foam was impregnated with an aqueous solution of hexachloroplatinic
acid with further treatment resulting in formation of platinum nano-
particles. Pt/γ-alumina/α-alumina foams were applied in hydrogenation
of EBF in this work to the mixture of mandelates without addition of any
optical modifier.
Deposition of γ-alumina on the surface of
α-alumina carrier was
performed in two ways: coating with slurry and deposition from hot
water solution of aluminium nitrate.
2. Experimental
The slurries were prepared by the following procedure: 5.0 g of
γ-alumina was mixed with 8–16 mL of 5% HCl aqueous solution with
subsequent ball milling for 30 min.
2.1. Preparation of ceramic foams, washcoating and deposition of the
active phase
The coating process was done similarly to the procedure for coating
PU sponges with
the calcined
α-alumina. However, no squeezing step was included as
Two types of commercial PU sponges (15 and 20 PPI where PPI
stands for pores per inch) were supplied by Recticel Oy. To remove
impurities, PU sponges shaped in a form of a cylinder with defined di-
mensions (length = 20 mm, diameter = 16 or 20 mm) were cleaned first
with distilled water and acetone and thereafter air-blown prior to use. A
clean PU sponge was then pretreated to create a sufficient surface
roughness for a better suspension adhesion [27] by being placed into
1 M NaOH (Merck, CAS number [1310ꢀ 73-2]) solution for 24 h, after
α
-alumina foams are not anymore flexible. The sample was
immersed once in the slurry, air-blown for removing the excess slurry
and dried.
An alternative washcoating process consisted of immersion of
α
-alumina samples in a aluminium nitrate solution to introduce
γ-alumina on the foam surfaces. Two immersions in 1 or 2 M aluminium
nitrate solution for 1 min each at 80 ◦C were performed.
2

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
In both cases, calcination was done at 600 ◦C for 6 h in an electrically
heated furnace. The heating schedule is shown in Fig. 1b illustrating that
the temperature was slowly increased from room temperature to 600 ◦C.
The same heating protocol was applied as mentioned above (first by
heating at a rate 50 ◦C/h to 300 ◦C and then to 600 ◦C with a ramp 5 ◦C/
min). Cooling down was also slow (5 ◦C/min) in order to avoid any
thermal stresses in the foam.
Combination of the two washcoating methods described above by
first applying γ-alumina coating from the corresponding suspension
followed by deposition from aluminium nitrate was also studied.
Deposition of platinum as the active catalytic component was done
using the following procedure: a certain amount of hexachloroplatinic
acid (Sigma Aldrich, 241ꢀ 010-7) (calculated as 5% wt. of γ-alumina in
the sample) was dissolved in 600 mL of water. The solution was pumped
through the samples of foams by a peristaltic pump (at the speed of
75 mL/min) for 24 h to ensure an efficient deposition of the platinum
complex on the foam surface. Deposition of platinum was monitored by
a colour change of the solution.
Fig. 2. Exemplary visualisation of measured distributions of local attenuation
coefficients using microfocus X-ray CT.
After the deposition, the foam was washed with ammonia (25 wt.%)
at room temperature followed by drying at 80 ◦C. The washed samples
were calcined at 400 ◦C for 3 h. Reduction with flowing hydrogen was
performed at 250 ◦C for 3 h. After reduction the samples were dried at
room temperature for 10 h.
diameter) was determined in the vertical positions.
Transmission electron microscopy (TEM) and X-ray diffraction were
applied for catalyst characterization. An energy-filtered transmission
electron microscope (EFTEM LEO 912 OMEGA, 120 kV) and a scanning
electron microscope (SEM/EDS, Jeol JSM-6400) were used to study the
microstructure of the support as well the distribution and the size of the
catalyst particles.
2.2. Characterization of the support and the catalysts
The crystal structure of the catalysts was determined by using X-ray
diffraction (SiemensD5000 with CuKa radiation).
The specific surface area and the pore volume were obtained by ni-
trogen physisorption using Sorptomatic 1900, Carlo Erba Instruments.
First the samples were outgassed under vacuum at 150 ◦C for 3 h and the
adsorption/desorption steps were carried at 77 K, using liquid nitrogen
as a coolant. The data were interpreted with the Brunauer-Emmett-
Teller isotherm and the t-plot methods.
The three-dimensional (3D) structures of the foams were analysed
using a proprietary microfocus cone-beam X-ray computed tomography
(μCT) setup [29] comprising a microfocus X-ray source (XrayWORX
XWT-190-TC), a two-dimensional (2D) flat panel X-ray image detector
(PerkinElmer XRD 0822 AP3 IND) and a motorized precision rotary
Catalyst images as well as elemental analysis data were obtained
with a Leo Gemini 1530 scanning electron microscope (SEM) equipped
with a ThermoNORAN + Vantage X-ray detector for energy dispersive
X-ray analysis (EDXA) analysis. The images were taken using the sec-
ondary electron and backscattered electron detectors at 15 kV, and the
in-lens secondary electron detector at 2.70 kV.
stage (FEINMESS DT 105-LM). According to Beer-Lamber’s law the in-
{
tensity measured at each detector pixel gets attenuated by exp
ꢀ
∫
}
μ
(r)dr as the X-ray photons traverse along ray path r through the
sample, with
μ
(r) being the sample’s local linear attenuation coefficient.
At an X-ray source voltage of 125 kV sets of one thousand 2D X-ray
projection images were acquired while each sample was rotated around
The macroporosity of the PU and alumina foams at different prepa-
ration steps (i.e. pristine PU, after pretreatment with NaOH, deposition
of a silica layer, a-alumina, after washcoating with g-alumina, as well as
after impregnation with hexachloroplatinic acid) was evaluated using an
epoxy casting method. About 25 mL of the resin (Elichem Resins Ltd)
and 1 mL of the catalyst were mixed and stirred manually for few mi-
nutes and then poured in the container until the whole sample was
covered. Afterwards, air inside the epoxy resin was removed by a vac-
uum pump. After drying for 24 h at room temperature, the samples were
cut out from the middle and polished to get a smooth surface. Finally,
the samples were washed with ultra-pure water and coated with carbon
for a better conductivity. SEM images were taken slide by slide and they
were attached to each other to make a panorama.
360◦. Local attenuation coefficient distributions
μ(x, y, z) are recon-
structed from the projection images using a proprietary implementation
of the Feldkamp-Davis-Kress algorithm [30,31] onto three-dimensional
voxel grids with a resolution of 30 μm voxel size, as exemplarily shown
in Fig. 2. Grey-scale colors represent local attenuation with dark grey
corresponding to strong attenuation (i.e. high density) and with light
grey corresponding to weak attenuation (i.e. low density). A threshold
was applied for the 3D visualisation.
2.3. Catalytic reaction
Hydrogenation of ethyl benzoylformate (EBF) to ethyl mandelate
(Fig. 3a) over macroporous Pt/Al2O3 catalyst in solvent mixture
comprising hexane/2-propanol (90/10) v/v was chosen as a reaction for
showing the catalytic activity of the foams. This reaction giving a
racemate of R and S mandelate has been previously investigated [26]
with the same solvent mixture to facilitate downstream chromato-
graphic separation with a chiral column.
The macroporosity (P) of the foams was calculated using the
following equation:
ρt
ꢀ
ρb
P =
∗ 100%
(1)
ρt
where ρt is the theoretical density (3980 kg/m3) and ρb is the bulk
density of the alumina foam.
The catalytic activity measurements were performed in an up-flow
fixed bed reactor (47 cm length and 3 cm internal diameter) at atmo-
spheric pressure under a flow of molecular H2. (Fig. 3a).
Compressive strength data of the manufactured carriers was ob-
tained using crush testing. L&W crush tester with two parallel plates (SE
048, Lorentzen & Wettre, Sweden) was used to detect the force needed
for an extrudate to collapse. Two plates were moved towards each other
using a hydraulic device, recording the pressure at which the catalyst
extrudates were broken. The moving speed of the plates was 1 mm/min.
The mechanical strength of the foams (20 mm height and 16 mm in
The testing of the porous catalysts was conducted at 25 ◦C. To avoid
interactions between the catalyst and oxygen, the reaction medium was
bubbled with Ar for 10 min before putting it in contact with the catalyst.
The liquid phase volume and the initial concentration of EBF were 0.9 L
and 5.6 mmol/L respectively. Typically, the experiments were carried
3

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Fig. 3. Reaction scheme for hydrogenation of EBF: a) reaction scheme, b) Up-flow fixed bed reactor.
3. Results and discussion
3.1. Synthesis of foams
The influence of various processing parameters on properties of
porous alumina foams was studied to optimize pore interconnectivity,
total porosity and mechanical strength.
One of the most important processes during the ceramic foam
preparation is the pretreatment of the PU sponge. The mechanical
strength of the foam depends on many parameters and one of them is the
attachment of the slurry to the sponge walls. Treatment with 1 M NaOH
resulted in cracks on the surface layer of the sponge (Figs. 4 and 5).
Based on our preliminary experiments such treatment was needed,
otherwise preventing adhesion of the silica layer and eventually for-
mation of stable α-alumina foams.
Etching with NaOH enhances adhesion of the coating to the sponge
walls by influencing the sponge surface layers.
Fig. 4. Surface of PU sponge before pretreatment.
Immersing in the silica suspension was done to cover the surface of
the polyurethane matrix with a silica layer enhancing the adhesion of
the slurry to the PU surface (Fig. 6). Fig. 6 illustrates that immersing in
the silica solution results in covering of the surface of the polyurethane
matrix with a silica layer.
The slurry for the production of the foams comprised water, grained
ceramic powders and an additive. To prepare a slurry with a set of
certain properties, it is necessary to change the slurry viscosity, pH or to
use different types of additives.
The liquid-to-solid ratio of the slurry plays an important role in the
preparation of
α-alumina carriers. Slurries with a solid content
exceeding 60 % are viscous and cover the surface of PU sponges with
thick and irregular layers. As a result, after dipping PU in a slurry, a large
part of the cells became blocked. The solid-liquid content was chosen in
a way, that the open porosity of the carrier was kept constant connecting
the whole foam volume.
Corundum powder with titania and magnesia additives was ground
in a planetary mill to decrease the particle size and to improve adhesion
parameters. The exact composition is given in the experimental part.
After coating with the slurry, the sponges were dried and sintered.
The surface area and porosity of the catalyst based on ceramics depend
on the thermal treatment applied during the manufacture. To optimize
sintering, it is important to control the heating rate to maximize the final
density and minimize the grain growth of the particles.
Fig. 5. Surface of PU sponge after etching with NaOH.
out with co-current gas (50 mL/min) and liquid flows (1 mL/min).
The liquid samples were taken periodically and analysed by gas
chromatography (GC) using Varian 3300 chromatograph equipped with
a chiral column (Silica Chirasil-DEX; length 25 m, diameter 0.25 mm,
film thickness 0.25 μm). Helium was applied as a carrier gas with a split
Alumina can be sintered into ceramic monoliths at temperatures of
ca. 1700 ◦C. Sintering at lower temperatures is possible if the grain size
is reduced by prolonged milling. Based on the measurements reported in
[32], the milling time of 45 min was chosen allowing a decrease of the
ratio of 33. The flame ionization detector (FID) and injector tempera-
tures were 270 and 240 ◦C, respectively. The temperature program of
the GC was 120 ◦C (25 min)–20 ◦C/min–190 ◦C (6 min). The GC anal-
ysis was calibrated with ethyl benzoylformate (Aldrich, 95 %,
25,891ꢀ 1), (R)-ethyl mandelate (Aldrich, 99 %, 30,998ꢀ 2) and (S)-
ethyl mandelate (Aldrich, 99 %, 30,997ꢀ 4).
average particle size from 50 μm to 20 μm which is sufficient for sin-
tering at 1450ꢀ 1500 ◦C.
4

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Fig. 6. SEM micrograph of the sponge (ex-20 PU) surface a) prior and after immersing in the silica suspension at different magnification b) 30x, c) 100 × .
Fig. 7. Images of αꢀ alumina foams, (a): photo; and SEM images of the foam surface b) two dippings with magnesia, d = 16 mm, ex-15 PPI, magnification 30 and c)
two dippings with magnesia, titania and the antifoam agent, d = 16 mm, ex-15 PPI, magnification 100 d)hollow strut of the foam shown in Fig. 7c.
5

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Fig. 8. Surface microstructure of the ceramic foam after sintering (two dippings with magnesia, titania and the antifoam agent, d = 16 mm, ex-15 PPI) at different
magnification a) lkx and b) 5kx.
Fig. 7 displays a photograph of the sintered ceramic foam and SEM
Table 1
images illustrating that the ceramic powder has sintered to give a porous
Surface area of foams.
foam with hollow struts.
Sample
Foam (wt.% γ-alumina/wt.%
-alumina)
Surface area
(BET), m2/g
Surface area (t-
plot), m2/g
Immersion of the PU sponge with the slurry was done with the liquid
content increasing at every subsequent step: 30–40 % of PVA in the first
coating, 50 % in the second and 60–70 % in the third step. With such a
procedure, a good pore permeability was achieved with less than 5% of
the pores blocked according to visual evaluation. Some of the formed
larger round holes are supposed to be generated when gases from the PU
sponge matrix were released. Such holes may compromise the me-
chanical stability of the support.
α
1
2
3
4
5
6
γ-alumina (from Al(NO3)3)/
-alumina (2/98)
γ-alumina (from suspension)/
-alumina (5/95)
γ-alumina (from suspension)/
-alumina (13/87)
Pt/ γ-alumina (from suspension)/
-alumina (0.3/6/93.7)
Pt/ γ-alumina (from suspension)/
-alumina (0.5/10/89.5)
22
48
α
46
57
α
61
α
39
60
60
α
SEM images of the samples (Fig. 8) show strut surfaces where the
particles have partly grown together with sharp particle edges being still
visible. Such strut surfaces can accommodate an additional coating
layer.
56
α
γ-alumina (combination)/
284
Bdl*
α-alumina (29/71)
-alumina
7
*
α
After two or three immersions in the slurry, the samples became
capable to endure sintering without collapsing. More experimental work
is, however, needed to establish what would be the exact threshold in
alumina loading which allows to prevent sintering.
Below detection limit (no measureable nitrogen adsorption for 0.5 of the
foam).
In this work, different ways of sintering and pre-sintering treatments
were investigated, with one of them showing a possibility to reduce the
amount of the slurry deposited on the PU sponge surface. Pre-sintering
at 300 ◦C for a sufficiently long time was necessary to obtain samples
with a required strength, as otherwise materials not exposed to such
treatment were not able to withstand sintering without critical struc-
tural damages. A need for additional pre-sintering can be explained by
diffusion in-between the ceramic slurry and molten polyurethane (at
300 ◦C), which leads to less damage of the structure of the formed
ceramic layer even polyurethane is decomposed caused by a rapid
temperature increase [33]. Mechanical properties of such materials,
the
basically the same procedure as for the foam manufacturing via the
replica technique. First -alumina foams were immersed into the slurry
with γ-alumina and then calcined in order to attach a layer of γ-alumina
α-alumina support with an additional layer of γ-alumina utilizing
α
on
α-alumina carrier. One or two immersions into the slurry were
applied. This method allowed to add up to 30 wt.% of γ-alumina
(determined by weighing) onto the foam walls. Apparently, washcoating
leads to a decrease in the volume of the open pores. Heat treatment at
600 ◦C was performed after the washcoating.
The second method to increase the surface area of the foams con-
sisted of immersing α-alumina foams or samples coated with γ-alumina
suspensions in a hot (75ꢀ 85 ◦C) solution of aluminium nitrate, resulting
in deposition of γ-alumina. Two immersions for 1 min each were suffi-
cient to increase the mass of the sample by ca. 5 wt.%. Sintering at
600 ◦C was applied after the washcoating giving the final foam.
however, could be improved during subsequent coating with α-alumina
or washcoating with γ-alumina.
Two additional procedures were involved to increase the surface
area by washcoating of the
α-alumina foams with γ-alumina.
The first method of increasing the surface area consisted of coating
SEM images of the foam obtained after washcoating
α-alumina
Fig. 9. Sintered γ-alumina /α-alumina foam prepared by washcoating α
-alumina foam using γ-alumina suspension, sintering at 600 ◦C, washcoating with aluminium
nitrate solution and calcination at 600 ◦C: a) photo and b) SEM image at magnifications b) 30x and c) 100x (three dippings with magnesia, titania and the antifoam
agent, d = 16 mm, ex-15 PPI).
6

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Fig. 10. SEM of a) sintered γ-alumina /
α
-alumina foam prepared by washcoating
α
-alumina foam using γ-alumina suspension, sintering at 600 ◦C, washcoating with
aluminium nitrate solution and calcination at 600 ◦C (three dippings with magnesia, titania and the antifoam agent, d = 16 mm, ex-15 PPI) and b) the corresponding
α
-alumina foam. Both images at magnifications 5kx.
Fig. 11. High resolution X-ray computed tomography of foams, vertical slice images and corresponding histograms of measured attenuation coefficients of a)
-Al2O3 and b) γ -Al2O3/
material (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).
α
α
-Al2O3. The histograms (grey shaded) have been empirically decomposed into contributions from void (blue) and ceramic (red and orange)
support with either aluminum nitrate or γ-alumina (not shown) exhibit
SEM images of the foam obtained by coating
α
-alumina with
morphology to
α-alumina support with somewhat less homogeneity of
γ-alumina and the foam obtained after washcoating the same support
with aluminium nitrate and subsequent calcination showed similar
morphology. These images (not shown) did not reveal holes that were
the particles and a minor decrease in dimensions.
As mentioned in the experimental section a combination of the two
washcoating methods was also applied meaning that coating with
γ-alumina was followed by deposition of aluminium nitrate. The SEM
image of such foam is presented in Fig. 9. Physico-chemical properties of
the wash-coated foams will be discussed below.
visible on the surfaces of α-alumina samples.
Surface of the foam prepared with washcoating first the α-alumina
foam with γ-alumina suspension and subsequently with aluminium ni-
trate given in Fig. 10 is shown at a higher magnification than in Fig. 9. In
these images the surfaces look much more extended and rough
compared to the smooth surface of
magnification.
α-alumina foam taken at the same
3.2. Physico-chemical and catalytic properties of foams
The high resolution X-ray computed tomography of foams (Fig. 11)
clearly illustrates that the cross- sections of the foams had decreased
The surface areas of different foams after washcoating was measured
using nitrogen adsorption. The specific surface areas of the materials are
summarized in Table 1.
after washcoating, because γ -Al2O3 fills the pores of the underlying
α
-Al2O3 structure. The greyscale values in the images represent recon-
structed X-ray attenuation, which is – for the considered ceramic com-
pound – proportional to their density. For the specific sample presented
in Fig. 11b, the density of γ -Al2O3 appears to be only 40 % of the density
The table reveals that the surface area of the foams after deposition
of γ-alumina has increased substantially. Typically, the surface area of
α
α
-alumina is 1ꢀ 5 m2/g. In the current study the surface area of
-alumina foam was probably even low as it could not be reliably
of
α -Al2O3, while the overall mass fraction per volume as measured by
determined from nitrogen physisorption. Such increase as expected was
more prominent at a higher washcoating load. Table 1 also illustrates
that the combined method, allowing a large amount of washcoat,
resulted in a substantial increase of the surface area (compare entries 3
and 6). The deposition of Pt did not markedly influence the surface area
(compare entries 3 and 5).
microfocus X-ray CT increased by factor 1.089 after washcoating.
More quantitative analysis of washcoating was performed by SEM
(Fig. 12).
In particular, the SEM images of the cross-section of the foams were
used to investigate differences in macroporosity. Fig. 12 shows cross-
7

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Fig. 12. Cross-sections of foams with increasing number of coating steps from a) 0 to e) 4. In the preparation two dippings with magnesia, titania and an antifoam
agent were done, d = 16 mm, ex-20 PPI.
sections of foams, which underwent different number of coatings with γ
Table 2
-alumina. The PU foams used for preparation of ceramic foams had a
Crushing strength of foams (16 × 20 mm) with a different number of coating
height of 20 mm being 16 mm in diameter.
steps from aluminium nitrate. In the preparation two dippings with magnesia,
Data for the same five
α
-alumina foams with different content of
titania and an antifoam agent were done, d = 16 mm, ex-20 PPI.
γ-alumina are shown in Table 2. The first entry corresponds to
α-alumina
Number of
coating
steps
Macroporosity
Weight,
Maximum
Vertical
foam without any washcoating. As a comparison, it can be mentioned
that a commercial sample of γ-alumina (BDH Ltd) has the compression
strength of 19.6 MPa, much exceeding the values for foams. At the same
time extrudates [33] made from H-beta zeolite with 30 % bentonite as a
binder gave a compression strength of 3ꢀ 5 MPa depending on the
preparation method which is in line with the current results.
g
pressure, kN
compression
strength, MPa
0
1
2
3
4
93
90
87
84
82
1.12
1.51
2.1
0.17
0.19
0.41
0.57
1.01
0.84
0.96
1.31
1.80
3.23
3.91
4.46
Table 2 also illustrates that foams with more coatings had a higher
8

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Fig. 13. Dependence of the crushing strength on the washcoating weight in samples with a different number of coating steps, a) maximum pressure, kN, b) vertical
compression strength. In the preparation two dippings with magnesia, titania and an antifoam agent were done, d = 16 mm, ex-20 PPI.
Fig. 14. Images of crushed γ-alumina (from suspension)/α-alumina: presumably α -alumina (a and b), presumably γꢀ alumina (c and d).
Fig. 15. 5% Pt/γ-Al2O3/α-Al2O3 foam a) photo, b) TEM image of Pt nanoparticles formed after reduction and calcination, c) Pt size distribution.
mechanical strength at the expense of lower pore volume and inter-
connectivity of the macropores, also visible in Fig. 12. An almost linear
dependence of the strength as a function of the solid load can be seen in
Fig. 13.
TEM images of the crushed γ-alumina (from suspension)/α-alumina
(Fig. 14) underline morphology and crystalline composition of the
achieved ceramic material.
Most of the particles have smooth surfaces and sharp edges, while a
An increase of the mechanical strength of ceramic supports, critical
for mechanical durability was previously reported for structured cata-
lysts after washcoating with γ-alumina [34].
part of the particles exhibit rough surfaces which can point on α
-alumina (Fig. 14 a and b) and γꢀ alumina (Fig. 14 and d).
TEM of the catalyst containing 5% Pt/γ-Al2O3/α-Al2O3 (entry 6 in
9

V. Shumilov et al.
Catalysis Today xxx (xxxx) xxx
Table 1) displays (Fig. 15) that the metal is homogeneously distributed
across the surface. The average size of Pt clusters was ca. 3 nm. The EDX
analysis showed that the metal loading was close to 5 wt% (more pre-
cisely 4.96–5.83 depending on the areas in the sample).
the work reported in this paper.
Acknowledgement
For the later catalyst applied in hydrogenation of ethylbenzoyl
formate a steady-state was achieved after 2 h. For the liquid sample
taken 3.5 h with time-on-stream reactant conversion was 81 % with an
overall selectivity to the corresponding alcohols (R-, S-mandelates)
equal to 87 %. The rest were by-products in small amounts giving the
overall mass balance closure exceeding 95 %.
The authors are grateful to Linus Silvander for SEM analysis.
Financial support from the Graduate School in Chemical Engineering
(GSCE) to Vladimir Shumilov and Academy of Finland to Tapio Salmi is
gratefully acknowledged.
References
[1] C.P. Stemmet, Gas-liquid Solid Foam Reactors: Hydrodynamics and Mass Transfer,
PhD Thesis, Technical University of Eindhoven, 2008, https://doi.org/10.6100/
IR635735.
4. Conclusions
The replica technique was used for preparation of macroporous
alumina carriers for catalysis. Pretreatment of the polyurethane (PU)
sponge was found to be critical. Treatment with NaOH resulting in
cracks on the surface layer of the sponge was found to be essential for
[2] M. Saber, T.T. Huu, C. Pham-Huu, D. Edouard, Chem. Eng. J. 185-186 (2012)
294–299.
[3] V. Blachou, D. Goula, C. Philippopoulos, Ind. Eng. Chem. Res. 31 (1992) 364–369.
[4] Y.S. Han, J.B. Li, Y.J. Chen, Q.M. Wei, J. Mater. Proc. Tech. 128 (2002) 313–317.
[5] J. Luyten, S. Mullens, J. Cooymans, A.M. De Wilde, I. Thijs, R. Kemps, J. Eur.
Ceram. Soc. 29 (2009) 829–832.
stable α-alumina foams formation. A good pore permeability was ach-
[6] S. Sharafat, N. Ghoniem, M. Sawan, A. Ying, B. Williams, J. Am. Ceram. Soc. 81
(2006) 455–460.
ieved by immersion of the PU sponge in the suspension several times
with increasing polyvinylalcohol content. with less than 5% of the pores
blocked according to visual evaluation.
[7] S. Dhara, P. Bhagarva, J. Am. Ceram. Soc. 86 (2003) 1645–1650.
[8] H.X. Peng, Z. Fan, J.R.G. Evans, J.J.C. Busfield, J. Eur. Ceram. Soc. 20 (2000)
807–813.
Sintering at 1450ꢀ 1500 ◦C into ceramic porous structures was done
using pre-milled particles. An efficient way of reduce the amount of the
slurry deposited on the PU sponge without structural damage was by
pre-sintering at 300 ◦C for a sufficiently long time.
[9] J. Saggio-Woyansky, C.E. Scott, W.P. Minnear, Am. Ceram. Soc. Bull. 71 (1999)
1674–1682.
[10] X. Zhu, D. Jiang, S. Tan, Mater. Res. Bull. 37 (2002) 541–553.
[11] P. Sepulveda, Ceram. Bull. 76 (1997) 61–65.
[12] H. Schmidt, D. Koch, G. Grathwohl, P. Colombo, J. Am. Ceram. Soc. 84 (2001)
2252–2255.
The surface area of
α
-alumina foams can be significantly enlarged by
-alumina foams into
coating with γ-alumina using either immersion of
α
[13] O. Lyckfildt, J.M.F. Ferreira, J. Eur. Ceram. Soc. 18 (1998) 131–140.
[14] W.P. Minnear, Processing of foamed ceramics, in: M.J. Cima (Ed.), Ceramic
Transactions, Forming Science and Technology for Ceramics, Am. Ceram. Soc.,
Westerville, OH, 1992, pp. 149–156.
γ-alumina suspensions followed by calcination or such foams or samples
coated with γ-alumina suspensions in a hot (75ꢀ 85 ◦C) solution of
aluminium nitrate, resulting in deposition of γ-alumina. The latter
method was considered to the most promising in terms of the surface
area and mechanical stability. Foams with more coatings of γ-alumina
had a higher mechanical strength at the expense of lower pore volume
and interconnectivity of the macropores. The strength was significantly
improved almost linearly when the loading of γ-alumina was increased.
[15] K. Ishizaki, S. Komarneni, Porous Materials: Process Technology and Applications,
Klumer Academic Publishers, London U.K, 1998.
[16] M.V. Twigg, J.T. Richardson, Ind. Eng. Chem. Res. 46 (2007) 4166–4177.
[17] B. Simon, M. Seeber, U.T. Gonzenbach, L.J. Gauckler, J. Mater. Res. 28 (2013)
2281–2287.
[18] A. Manonukul, P. Srikudvien, M. Tange, C. Puncre, Mater. Sci. Eng. A. 655 (2016)
388–395.
[19] L. Samain, A. Jaworski, M. Eden, D.M. Ladd, D.-K. Seo, F.J. Garcia-Garcia,
U. Haussermann, J. Solid. St. Chem. 217 (2014) 1–8.
Homogeneous distribution of Pt in the 5% Pt/γ-Al2O3/α-Al2O3
[20] E. Behravesh, L. Hupa, T. Salmi, D.Yu. Murzin, Catal.-Spec. Period. Rep., Royal Soc.
Chem. 28 (2016) 28–50.
catalyst containing 5 wt.% of Pt with the average size of Pt clusters of ca.
3 nm was achieved using the impregnation method.
[21] M. Twigg, J.T. Richardson, Stud.Surf. Sci. Catalysis 162 (2006) 135–142.
[22] R. Faure, F. Rossignol, T. Chartier, C. Bonhomme, A.S. Maitre, G. Etchegoyen,
P. Del Gallo, D. Gary, J. Europ. Ceram. Soc. 31 (2011) 303–312.
[23] V.N. Grunskii, M.G. Davidkhanova, M.D. Gasparyan, S.E. Zolotukhin, Fibre Chem.
51 (2019) 240–243.
Catalytic activity was tested in hydrogenation of ethyl benzoylfo1-
mate. The catalyst was found to be active demonstrating 87 % selectivity
to the desired alcohol products at 81
benzoylformate.
% conversion of ethyl
[24] O.V. Starodubtseva, V.N. Grunskii, A.V. Bespalov, G.V. Avramenko, A.V. Ignatov,
Yu.A. Pinchuk, Khimich. Tekhnol. (Moscow) 14 (2013) 599–605.
[25] A.V. Ignatov, A.E. Varakutin, I.N. Solov’eva, I.B. Karmanova, I.A. Kozlov, M.
N. Semenova, V.V. Semenov, Russ.Chem. Bull. 67 (2018) 1394–1400.
[26] G. Martin, P. Maki-Arvela, D.Yu. Murzin, T. Salmi, Top. Catal. 57 (2013)
1576–1581.
CRediT authorship contribution statement
Vladimir Shumilov: Investigation, Data curation. Alexey Kirilin:
Investigation. Anton Tokarev: Investigation. Stephan Boden: Investi-
gation, Data curation. Markus Schubert: Investigation, Data curation.
Uwe Hampel: Methodology, Writing - review & editing. Leena Hupa:
Project administration, Conceptualization. Tapio Salmi: Project
administration, Conceptualization. Dmitry Yu. Murzin: Supervision,
Conceptualization, Writing - original draft, Writing - review & editing.
[27] H.E. Bergna, W.O. Roberts, Colloidal Silica: Fundamentals and Applications, CRC
Press, Baton Rouge, 2005.
[28] L. Jiao, H. Xiao, Q. Wang, J. Sun, Polym. Degrad. Stab. 98 (2013) 2687–2696.
[29] S. Boden, M. Bieberle, G. Weickert, U. Hampel, Powder Techn. 188 (2008) 81–88.
[30] L.A. Feldkamp, L.C. Davis, J.W. Kress, J. Opt. Soc. Amer. A 1 (1984) (1984)
612–619.
[31] A.C. Kak, M. Slaney, Principles of Computerized Tomographic Imaging, IEEE Press,
New York, 1987. ISBN:0-87942-198-3.
[32] T. Fey, U. Betke, S. Rannabauer, M. Scheffler, Adv. Eng. Mat. 19 (2017), 1700369.
[33] Z. Vajglova, N. Kumar, M. Peurla, L. Hupa, K. Semikin, D. Sladkovskiy, D.
Yu. Murzin, Ind. Eng. Chem. Res. 58 (2019) 10875–10885.
[34] A. Cybulski, J.A. Moulijn, Structured Catalysts and Reactors, CRC Press, Boca
Raton, FL, 2005.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
10